Carbon Robotics Launches Large Plant Model Trained on 150 Million Plants

"When our robots can understand any plant in any field immediately and adapt behavior in real-time, farmers immediately get maximum value from the machines," Paul Mikesell, founder and CEO of Carbon Robotics, said in the company's announcement on Monday.

Carbon Robotics just launched its Large Plant Model (LPM), an AI system trained on 150 million labeled plants collected from over 100 farms across 15 countries. This marks the first practical solution to agriculture's biggest automation problem.

Until now, every time a farmer encountered a new weed variety, Carbon Robotics' LaserWeeder fleet required 24 hours of retraining. The company had to create new data labels, retrain the entire machine, and wait. In agriculture, where pest pressure compounds daily and growing windows are measured in weeks, 24 hours of friction is an eternity. This constraint was the reason most agricultural AI systems never scaled beyond controlled environments.

The LPM eliminates that friction entirely. A farmer can now point at a plant and tell the machine, "This is what I want you to kill." No retraining or waiting. The model understands plants at a structural and biological level. When it encounters a plant it's never seen before, it reasons about what it's looking at based on deep pattern recognition across 150 million training examples.